Skip to content

op.py refactor & improved TP support #6

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 15 commits into
base: production
Choose a base branch
from

Conversation

a-alveyblanc
Copy link
Collaborator

@a-alveyblanc a-alveyblanc commented Feb 25, 2025

This is a large set of changes, so here are some patch notes:

Changes from refactor:

  • Construct all operators using sufficiently accurate quadrature
  • Separate operator construction (matrices.py) from operator application (op.py)
  • Add utilities to support new operator construction methods
    • Getting basis, nodes, quadrature rules, etc. based on discretization type
  • Remove FirstAxisIsElementsTag usage in op.py
    • This can be reversed if necessary, but it's a fate that was decided for upstream grudge

Changes related to TP elements:

  • Add algebraic DAG transformations for fast operator evaluation with TP elements
  • Add subclass of FusionContractorArrayContext with TP DAG transformations
    • use_tp_transforms kwarg to toggle on and off based on discretization type
  • Add toggle for fast operator evaluation (defaults to on unless overintegration is requested)

Overintegration will require a slight fix to modepy: inducer/modepy#111

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant